A problem in the art of peak detectors is that they use capacitors to store the peak amplitude of the signal whose peak is being detected. It is desirable to store the peak value as long as needed in the capacitor, e.g., for use in thresholding applications, especially where there may be long periods of reception of only a single value in the received signal, e.g. a long string of zeros. The value stored on the capacitor is buffered when providing its output for further processing. Although such buffering is necessary, disadvantageously, when the peak detector circuit is manufactured using bipolar devices, such buffering causes the capacitor to discharge, thereby reducing the period of time that the capacitor accurately stores the peak value.